Date:         Tue, 14 Nov 2006 15:57:14 -0500
Reply-To:     Doug Rohde <drohde01@COMCAST.NET>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Doug Rohde <drohde01@COMCAST.NET>
Subject:      Re: Creating multiple datasets from one dataset based on value of

I may be a little late getting into the game on this one, but Mike, I
think you are setting yourself up for a world of pain that will only grow
as your number of datasets increases.  Even worse, you are setting up
whoever comes after you for even more pain.

Let's say you want to make a structural change to the data sets - Instead
of doing it once, now you have to write macro code to handle each dataset
individually.  And then you have to write even more macro code to verify
what you did was correct.  What if you want to look at data across
companies?

So far, the only reason you've given for doing this is the vague "It works
better for me," which still tells us nothing about your situation.  The
only valid reason I can think of for doing this is performance, but unless
you are running SAS on an Atari 2600, a couple million rows should not
pose any serious performance issues.

But hey, you are free to ignore me and Toby and the rest.  Good luck.

Doug R

On Tue, 14 Nov 2006 12:24:54 -0800, MikeInWV <mikeslover@GMAIL.COM> wrote:

>Thanks to Mark and & Ya for their code.  I made some modifications to
>Ya's code in particular and worked extremely well.  Luckily for me the
>data was pretty easy to sort!  On a subset of data it created 50 some
>permanent datasets just like I want.
>
>Mark- thanks for your understanding.  Toby- I do have a choice in not
>creating company level datasets but in my situation it works best to
>break it up into smaller pieces for the inevitable questions and edits.
> And if I'm lucky, I'll pick up another 200 companies in the next week!
>
>
>Cheers!